Persons with mental and substance use disorders account for a substantial and growing portion of Emergency Department (ED) visits in the United States. EDs frequently lack the capacity to provide behavioral health staff in EDs, pointing to the potential efficiencies and increased access to care that could be provided via centralized delivery of these services via telemedicine. However, little is known about the outcomes and costs of this approach for delivering behavioral services in ED settings. Beginning in 2009, the South Carolina Department of Mental Health developed and began implementing a statewide program under which psychiatrists are providing around-the- clock coverage for consultation with hospital emergency departments via high speed internet video connections. Preliminary data on service utilization for this model of care delivery indicate clinically substantial and statistically significant benefits on reducing length of ED stay, hospitalization, and total costs, and increasing rates of outpatient mental health follow-up. The findings point to a need for a rigorous assessment of quality, costs, and long-term sustainability of this approach. Grant funding for the program and the evaluation will end in 2012. This application proposes to conduct a rigorous economic analysis of this approach to restructuring delivery of behavioral care in ER settings. Analyses will draw on South Carolina's comprehensive health data warehouse, a unique, all-payer dataset that makes it possible to comprehensively track treatment contacts for individuals across the behavioral health, health, and social service systems. Data will be compared with patients from other EDs in South Carolina and, using Medicaid claims data, from four adjacent Southern states. The dataset will be enriched through reviews of electronic records from the interventionists and from an organizational survey of emergency rooms participating in the program. The study will provide data on cost, quality, and sustainability of this approach using econometric methods that can provide plausibly causal estimates of these impacts, to determine whether this organizational model should be scaled nationally.